Reservoir Saturation Prediction Using Multi-Property Petrophysical Parameters and Stacking Ensemble Learning
Abstract
tation can be unstable due to shale effects and multi-property coupling. This study uses multi-property petrophysical parametersincluding
density, porosity, true resistivity, and multi-frequency polarization attributesas inputs. Robust preprocessing (IQR-based clipping) and fea
ture standardization are applied, followed by feature construction and screening. Five base learners (SVR, Random Forest, Gradient Boosting,
XGBoost, and LightGBM) are trained and fused through stacking ensemble learning. The stacking model provides improved overall fitting
compared with individual models and simple fusion, and predictions are generally consistent with measured values. Independent testing indi
cates systematic deviations across saturation ranges, suggesting that sample distribution and limited endpoint samples affect generalization.
The proposed workflow enables rapid quantitative saturation prediction for intervals with limited core data or unstable conventional interpre
tation, supporting integrated reservoir evaluation and sweet-spot identification.
Keywords
Full Text:
PDFReferences
[1] Archie G E. The electrical resistivity log as an aid in determining some reservoir characteristics[J]. Transactions of the AIME, 1942,
146(01): 54-62.
[2] Waxman M H, Smits L J M. Electrical conductivities in oil-bearing shaly sands[J]. Society of Petroleum Engineers Journal, 1968, 8(02):
107-122.
[3] Wolpert D H. Stacked generalization[J].Neural Networks, 1992, 5(2): 241-259.
[4] Spearman C. The proof and measurement of association between two things[J]. The American Journal of Psychology, 1904, 15(1): 72-
101.
[5] Tukey J W. Exploratory data analysis[M]. Reading, MA:Addison-Wesley, 1977.
[6] Friedman J H. Greedy function approximation:A gradient boosting machine[J]. The Annals of Statistics, 2001, 29(5): 1189-1232.
[7] Breiman L. Random forests[J]. Machine Learning, 2001, 45(1): 5-32.
[8] Smola A J, SchlkopfB.A tutorial on support vector regression[J]. Statistics and Computing, 2004, 14: 199-222.
[9] Chen T, Guestrin C. XGBoost:A scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on
Knowledge Discovery and Data Mining.New York:ACM, 2016: 785-794.
[10] Ke G, Meng Q, Finley T, et al. LightGBM: a highly efficient gradient boosting decision tree[C]//Advances in Neural Information
Processing Systems 30 (NeurIPS 2017). Red Hook,NY: Curran Associates, 2017: 3146-3154.
DOI: http://dx.doi.org/10.70711/itr.v3i3.9237
Refbacks
- There are currently no refbacks.
